Tutorial Overview
Series index: Hermes Agent Tutorial Series
This tutorial covers Hermes Agent’s Terminal Interface — the CLI commands and TUI features for users who prefer terminal interaction.
Note: If you deployed Hermes via Messaging Gateway (Tutorial 4), this article provides terminal basics. If you’re here for terminal-only usage, this is your primary interface guide.
What you will learn
- ✅ CLI entry points and their purposes
- ✅ TUI features: multiline editing, autocomplete, history
- ✅ Key slash commands and their usage
- ✅ Personalities system for different contexts
- ✅ Session management and persistence
CLI Entry Points
Main commands
Hermes provides several CLI commands for different purposes:
| Command | Purpose | Usage |
|---|---|---|
hermes |
Interactive TUI | Start conversation |
hermes gateway |
Messaging gateway | Deploy to platforms |
hermes model |
Model management | Switch providers |
hermes setup |
Configuration wizard | Initial setup |
hermes config |
Settings management | View/edit config |
hermes skills |
Skills management | List/create skills |
hermes memory |
Memory management | View/export memory |
hermes doctor |
Diagnostics | Troubleshoot issues |
Command flow
flowchart TD
A[hermes] --> B[Interactive TUI]
C[hermes gateway] --> D[Platform daemon]
E[hermes model] --> F[Provider switch]
G[hermes setup] --> H[Config wizard]
I[hermes doctor] --> J[Health check]
style B fill:#e1f5ff
style D fill:#fff3e0
TUI Features
Terminal User Interface
When you run hermes, you get a full TUI with:
flowchart LR
subgraph TUI
A[Header: Model + Usage]
B[Conversation area]
C[Input box]
D[Status line]
end
style A fill:#e8f5e9
style B fill:#e1f5ff
Multiline editing
Press Ctrl+J or Alt+Enter to create newlines in your message:
You: I need to create a function that:
1. Validates input
2. Processes data
3. Returns result
[Ctrl+J for new line]
Autocomplete
Hermes provides autocomplete for:
- Slash commands: Type
/and pressTab - Model names:
/model+Tab - Skills:
/skill+Tab
History navigation
Ctrl+P/Ctrl+N: Previous/Next messageCtrl+R: Search history↑/↓: Arrow keys also work
Interrupt handling
Ctrl+C: Interrupt current response- Send new message: Redirects agent mid-task
/stop: Graceful stop command
Key Commands
Essential slash commands
| Command | Purpose | Example |
|---|---|---|
/new |
Fresh conversation | /new |
/reset |
Clear context | /reset |
/model |
View/change model | /model openrouter:claude-3 |
/retry |
Regenerate last response | /retry |
/undo |
Undo last turn | /undo |
/compress |
Compress context | /compress |
/usage |
Show token usage | /usage |
/insights |
Session analytics | /insights --days 7 |
/skills |
Browse skills | /skills |
/stop |
Interrupt work | /stop |
/exit |
Exit Hermes | /exit |
Model switching
/model # View current model
/model openrouter:auto # Switch to auto-select
/model anthropic:claude-4 # Switch to Claude 4
/model openai:gpt-4o # Switch to GPT-4o
Memory commands
/memory show # View user profile
/memory list # List insights
/memory search "api design" # Search past sessions
Context management
/compress # Compress long context
/usage # Show current usage
/reset # Full context reset
Personalities System
What are personalities?
Personalities are preset configurations for different contexts:
flowchart TD
A[Base Hermes] --> B[Developer]
A --> C[Researcher]
A --> D[Writer]
A --> E[Assistant]
style A fill:#e1f5ff
Available personalities
| Personality | Focus | Ideal for |
|---|---|---|
developer |
Code, debugging | Programming tasks |
researcher |
Analysis, synthesis | Literature review |
writer |
Drafts, editing | Content creation |
assistant |
General help | Everyday tasks |
Switch personalities
/personality developer
/personality researcher
/personality default
Create custom personality
# Create personality file
nano ~/.hermes/personalities/my-work.md
Format:
---
name: my-work
description: Work-focused assistant
---
## Style
- Professional tone
- Bullet-point summaries
- Action-oriented recommendations
## Preferences
- Focus on actionable output
- Minimize explanation unless asked
Activate:
/personality my-work
Session Management
Session persistence
Hermes automatically saves sessions to:
~/.hermes/sessions/
├── 2026-04-13-001.json
├── 2026-04-13-002.json
└── ...
Session commands
| Command | Purpose |
|---|---|
/new |
Start new session |
/resume SESSION |
Resume past session |
/sessions |
List sessions |
Resume a session
/sessions # List recent sessions
/resume 2026-04-12-003 # Resume specific session
Session search
Find past conversations:
/search "authentication bug"
/search --from 2026-04-01 --to 2026-04-10
Platform-Specific Status
For CLI users
/status # Show daemon status
/platforms # Show connected platforms
Set home platform (for gateway users)
/sethome telegram # Telegram as primary
/sethome cli # CLI as primary
Configuration
View configuration
hermes config show
Set values
hermes config set model.default "openrouter:auto"
hermes config set memory.nudges.enabled true
hermes config set transcription.provider whisper
Edit config file
nano ~/.hermes/config.yaml
Example config:
model:
default: openrouter:auto
fallback: openrouter:claude-3-haiku
memory:
nudges:
enabled: true
interval: 10 # turns
gateway:
telegram:
enabled: true
token: "YOUR_TOKEN"
transcription:
provider: whisper
Troubleshooting
TUI not starting
Cause: Terminal compatibility issue.
Fix:
# Check terminal type
echo $TERM
# Force basic mode
hermes --mode simple
Commands not working
Cause: Slash command syntax wrong.
Fix: Commands start with / and no arguments need quotes:
/model openrouter:auto # Correct
/model "openrouter:auto" # Incorrect
Session not persisting
Cause: Session saving disabled.
Fix:
hermes config set sessions.save true
Summary
CLI/TUI is Hermes’s native interface:
- Multiple entry points —
hermes,hermes gateway,hermes model - Rich TUI — Multiline, autocomplete, history
- Slash commands —
/model,/memory,/skills - Personalities — Context-specific presets
- Session management — Auto-save, resume, search
Key takeaways
- ✅
hermesstarts interactive TUI - ✅ Tab completes commands and models
- ✅
/modelswitches providers instantly - ✅ Personalities customize behavior
- ✅ Sessions auto-persist and searchable
Series navigation:
- ← Previous: Tutorial 4: Messaging Gateway — Multi-Platform Integration
- → Next: Tutorial 6: Multi-Model Configuration — Power Configuration
- Back: Series Index
Decision Note: You now have two deployment options mastered. Messaging Gateway (Tutorial 4) and CLI/TUI (this article). Tutorial 6 applies to both — multi-model configuration enhances whichever interface you use.